On Thu, Jun 4, 2009 at 10:48 AM, HARRIS TONY <[log in to unmask]
> wrote:
> Try this search then.
>
> http://www.google.com/squared/search?q=walter+sickert+paintings
>
> Not much about the paintings really, mostly all about if he was Jack the
> Ripper or not. Seems like google is going for the tabloid angle on it's
> search programming. Also if the data is being presented as 'facts', well
> Walter Sickert + Jack the Ripper = fact? I don't think so.
Heh, it's full of errors, yes. But it's not terrible considering that it's
attempting to extract all this information from unstructured text.
You can also 'correct' it by removing rows, adding and removing columns and
changing values (which have a 'confidence' indicator associated with them, I
note).
Here's my quick attempt at getting a list of museums:
http://www.google.com/squared/table/ag_1u_7QBxu2H06QCffCGObA
There's only 41, which is a lot less than the 3,651 on Freebase (
http://www.freebase.com/view/architecture/museum), but it pulls out some
interesting ones, and from foreign languages too.
As Dan Z hints in his mw2009 paper, Google like to do things the top-down
way, gathering data by mining their huge indexes and the improving it
through the implicit and explicit feedback of their huge userbase.
At the same time, data is being collected the bottom-up way (although mostly
by users rather than institutions) through efforts like Wikipedia (&
DBpedia), Freebase, MusicBrainz, Open Library and so on.
I suspect the most successful projects will do a bit of both though.
Frankie
--
Frankie Roberto
Experience Designer, Rattle
0114 2706977
http://www.rattlecentral.com
Sent from Sheffield, United Kingdom
****************************************************************
For mcg information and to manage your subscription to the list,
visit the website at http://www.museumscomputergroup.org.uk
****************************************************************
|